tuieval 0.2.0.dev119__tar.gz → 0.2.0.dev121__tar.gz

{tuieval-0.2.0.dev119 → tuieval-0.2.0.dev121}/CHANGELOG.md

@@ -2,6 +2,8 @@
 
 ## Unreleased
 
+- Fix: Results → Per question opened the answers of the model one column to the right of the cell picked, and crashed on the last model's column.
+- Fix: `v` on the run screen switches to the last model still running when the stream shows one that has finished.
 - Setup opens with no models ticked: at launch, after New run, and after Back from a run or tune. Packs and options are still remembered.
 - Tuning: a server that refuses the `ignore_eos` request field is asked again without it, so strict servers can be tuned; the profile notes that workload answers may end early.
 - Fix: a server that crashes mid-request (the stream ends with no finish reason) was graded as an empty or cut-short answer; it's a server error now, never counted against the model.

{tuieval-0.2.0.dev119 → tuieval-0.2.0.dev121}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: tuieval
-Version: 0.2.0.dev119
+Version: 0.2.0.dev121
 Summary: Evaluate local and frontier LLMs on your own questions: accuracy, speed, tokens and PASS/FAIL verdicts, in the terminal.
 Project-URL: Homepage, https://github.com/ashe-wb/tuieval
 Project-URL: Issues, https://github.com/ashe-wb/tuieval/issues

{tuieval-0.2.0.dev119 → tuieval-0.2.0.dev121}/src/tuieval/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.2.0.dev119'
-__version_tuple__ = version_tuple = (0, 2, 0, 'dev119')
+__version__ = version = '0.2.0.dev121'
+__version_tuple__ = version_tuple = (0, 2, 0, 'dev121')
 
 __commit_id__ = commit_id = None

{tuieval-0.2.0.dev119 → tuieval-0.2.0.dev121}/src/tuieval/tui.py

@@ -2779,8 +2779,8 @@ class RunScreen(Screen):
     def action_watch_next(self):
         """Stream the next model running alongside (from its next answer)."""
         active = self.active()
-        if len(active) < 2:
-            self.notify("Only one model is running right now.")
+        if not active or active == [self.watch]:
+            self.notify("Only one model is running right now." if active else "No model is running right now.")
             return
         self.watch = active[(active.index(self.watch) + 1) % len(active)] if self.watch in active else active[0]
         with self.buf_lock:
@@ -3296,7 +3296,7 @@ class ResultsScreen(Screen):
         here = self.app.engine.machine().id
         table = self.query_one("#per-question", DataTable)
         table.clear(columns=True)
-        table.add_columns("Test", "Pack", "Level", "Added", "×tokens", *models)
+        table.add_columns(*self.PQ_FIXED, *models)
         self.pq_tests, self.pq_models = {}, models
         for t in tests:
             cells = []
@@ -3415,11 +3415,13 @@ class ResultsScreen(Screen):
         return (f"{mark} {tok_s} {secs}" + (" [red]cut[/red]" if c["truncated"] else "")
                 + (" [cyan]◐[/cyan]" if c["partial"] else ""))
 
+    PQ_FIXED = ("Test", "Pack", "Level", "Added", "×tokens")   # Per question's columns before the models
+
     @on(DataTable.CellSelected, "#per-question")
     def per_question_selected(self, event):
-        col = event.coordinate.column - 4
+        col = event.coordinate.column - len(self.PQ_FIXED)
         t = self.pq_tests.get(event.cell_key.row_key.value)
-        if not t or col < 0:
+        if not t or not 0 <= col < len(self.pq_models):
             self.notify("Pick a model's cell to open its answers.")
             return
         c = t["cells"].get(self.pq_models[col])

{tuieval-0.2.0.dev119 → tuieval-0.2.0.dev121}/tests/test_tuieval.py

@@ -847,6 +847,17 @@ class Parallel(unittest.TestCase):
         self.assertEqual({out["first"], out["switched"]}, {"a", "b"})   # v streams the other model
         self.assertTrue(any(r.get("ran_alongside") for r in self.results("a")["results"]))
 
+    def test_watch_next_moves_to_the_last_model_running(self):
+        from tuieval.tui import RunScreen
+        run = unittest.mock.MagicMock(watch="a", buf={"reasoning": [], "answer": []}, buf_lock=threading.Lock())
+        run.active.return_value = ["b"]          # a finished; only b is still running
+        RunScreen.action_watch_next(run)
+        self.assertEqual(run.watch, "b")
+        run.active.return_value = ["b"]
+        RunScreen.action_watch_next(run)          # already watching the only one
+        self.assertEqual(run.watch, "b")
+        run.notify.assert_called_with("Only one model is running right now.")
+
     def test_skip_one_model_keeps_the_other(self):
         e, events = self.engine()
         jobs = e.plan(["a", "b"], ["apps"], None, "certify", False, {})
@@ -1750,7 +1761,23 @@ class PTAIndex(unittest.TestCase):
                         await pilot.pause(0.3)
                         s = app.screen
                         sorts["opened"] = s.entries[s.index]["record"]["test"]
-                        print(json.dumps({"rows": rows, "bars": "P parsimony" in tri, "disagree": [both, alone],
+                        s.dismiss()
+                        await pilot.pause(0.2)
+                        app.pq_models = set()
+                        app.screen.query_one("#pq-disagree").value = False
+                        app.screen.query_one(TabbedContent).active = "tab-per-question"
+                        app.screen.fill_per_question()
+                        await pilot.pause(0.3)
+                        opened = {}
+                        pq.focus()
+                        for col in range(5, len(pq.columns)):                  # each model's cell opens its own answers
+                            pq.move_cursor(row=0, column=col)
+                            await pilot.press("enter")
+                            await pilot.pause(0.3)
+                            opened[str(pq.ordered_columns[col].label.plain)] = app.screen.entries[0]["model"]
+                            app.screen.dismiss()
+                            await pilot.pause(0.2)
+                        print(json.dumps({"rows": rows, "bars": "P parsimony" in tri, "disagree": [both, alone], "opened": opened,
                                           "tab": tab, "sorts": sorts, "fails": fails}))
                 asyncio.run(go())
             """)
@@ -1760,6 +1787,8 @@ class PTAIndex(unittest.TestCase):
             self.assertIsInstance(got, dict, got)
             self.assertTrue(got["bars"])
             self.assertEqual(got["disagree"], [4, 0])
+            self.assertGreaterEqual(len(got["opened"]), 2)
+            self.assertEqual(got["opened"], {m: m for m in got["opened"]})   # a cell opens its column's model
             self.assertEqual(got["tab"], "tab-pta")                  # Results opens on the PTA index
             self.assertEqual([(r[0], r[3]) for r in got["rows"]][0], ("good", "4/4"))
             s = got["sorts"]   # a click on a column name sorts by it, a second click reverses